Review Comment:
This manuscript was submitted as 'full paper' and should be reviewed along the usual dimensions for research contributions which include (1) originality, (2) significance of the results, and (3) quality of writing.
Most of the comments from my previous review are addressed in this new version of the paper. Especially the examples added helped me to understand the idea of the approach, this was a clear improvement.
The contribution of this paper, the implementation of a social internet of things framework for domotics, is explained and put into relation to other existing approaches. The authors show that their approach is able to solve the interoperability problem which often is present in AmI environments.
Nvertheless, there are still some points, I would like to see clarified. In particular:
Knowledge-based architecture (Chapter 3):
It is an improvement that you refer to Table 1 in the text and explain it later on. Nevertheless, this explanation is not close enough to the table in my opinion. I had some difficulties to find the items of the table in the text. It could for example help to use the exact names from the table in the text (e.g. use “Social agent” or “Object” in the text instead of “social object”).
Service-oriented architecture (Chapter 3):
Before you say that nodes, which are social objects, can make posts to a friend’s wall, it could be mentioned that being a social object includes the possibility of having friends.
Semantic matchmaking (Chapter 3): The examples you added really help to understand the approach, but there is still an open question:
Introducing example (starting at the end of the first column of page 5): R_1 is a subclass of S_1, so here seems to be something wrong with the example. Maybe S_1 is the class WasherDrier and R_1 is the intersection of WashingMachine and LargeCapacity?
Social entities and relationships (Chapter 3):
I have problems understanding Figure 1:
- What is ACK?
- Why does a node only get another device’s profile in the friends case (a), but not in the follower case (b), I thought that the decision to become a follower is based on the profile?
- Is the option to write on the wall reflected in case a of the figure?
Collaborative adaptivity (Chapter 3):
The concept “request” is still not clear to me: you say that “Each post contains all sensed perceptions and events observed by the social device and is considered as a request for system reconfiguration [...]” (quote from page 5). So my question here is: if this is already an observation reflecting “the world” why do I even need a process to “fulfil” the request? The process you describe seems to be some kind of bakward-reasoning which makes sure that for every observation there has at least one service that reacts to it. Is that how I need to understand it? Please clarify that in your text.
Chapter 4:
- I do not see why FullClose is selected, is the negation in Lamp_On the reason (page 12)?
- Figure 13: why does the figure contain “detectsIntrusion only IntrusionForLamp”, I see that because of the equivalence of the intrusion to the intersection of the two other intrusions that is one possible option, but “detectsIntrusion only Intrusion” would be a valid alternative in my opinion.
Minor remarks and typos:
Abstract:
maybe it would be helpful to better explain the problem with the current implementations and the advantage of your solution already in the abstract (you do it in the introduction, but in the abstract, you are rather vague).
Page 6:
Like in SNSs, in the framework proposed here object’s wall is the main channel for sharing knowledge. → the object’s wall
page 13: The WS selects the SC_2… → selects SC_2
page 14: As detailed in Table 3, KNX installation consisted… → the KNX installation
| 
